The first post in this series can be found here.
Last time I looked at a couple literature reviews about the methods of action of Plan B emergency contraception (levonorgestrel, LNG). This time I’m presenting On the the mechanisms of action of short-term levonorgestrel administration in emergency contraception (Durand, et al., 2001)
Experimental Design
Forty-five healthy, surgically sterilized women, aged 29 years to 35 years old (mean age 31 years), with regular menstrual cycles (cycle lengths between 25 days and 32 days) were recruited for this study. None had used hormonal contraception or any other medication within 6 months prior to the study. Participants were in good health as determined by medical history, physical examination, and routine screening laboratory tests, including Papanicolaou smear. Body weight, height, and blood pressure of each participant were registered by one investigator. Participants were issued a menstrual calendar on which they recorded details of all bleeding episodes throughout the study.
This bit tells us that were these women not surgically sterilized, we would have no reason to not expect that they were capable of both ovulation and, under the right circumstances, pregnancy.
The study was conducted in two consecutive cycles. Cycle was defined as the time elapsed from the first day of a spontaneous menstrual bleeding until the day preceding the next menses.
That is, cycles were defined symptomatically, not based on the mythical average 28-day cycle. Two cycles were tracked for each participant, control cycle in which no drug was administered and a treated cycle in which LNG was administered.
All participants were admitted during the first 10 days of their menstrual cycle. Before the control cycle, women were randomly allocated into three different groups as follows: Group A women received two doses of 0.75 mg LNG (Postinor, Gedeon Richter, Budapest, Hungary) taken 12 h apart, with the first dose given on the morning of day 10 of the menstrual cycle; Group B women received the same dose of LNG immediately after positive LH [luteinizing hormone] detection in urine; and Group C women received the same dose of LNG 48 h after positive detection of urinary LH.
- Group A received the drug on cycle day 10, regardless of when they ovulated.
- Group B received the drug after the LH surge, which shortly precedes ovulation.
- Group C received the drug 48 hours after the LH surge, i.e., presumably after ovulation has occurred.
During both cycles (control and treated), all women were asked to monitor urinary LH every morning, starting on the 11th day of the menstrual cycle until the presence of LH was detected. At this time, transvaginal ultrasound was performed daily until follicle rupture (FR) was observed. This was established by the presence of at least three of the following findings: acute decrease in mean diameter or disappearance of the follicle, presence of thickened irregular borders, increased echogenecity within the follicle, and presence of free intraperitoneal fluid [16]…
The LH surge was used as an indicator of imminent ovulation. Ovulation was observed and timed via ultrasound. The determination that ovulation had occurred was made by observing the follicle (which had contained the ovum) get smaller, disappear, thicken, or show up more brightly in the image, or the presence of fluid in the abdominal cavity.
Daily blood samples were obtained from the day of positive LH detection in urine until the day menses began…The main purpose for measuring serum LH was to precisely determine, rather than based only on urinary LH detection, the actual time at which LNG was administered during the menstrual cycle. Follicular phase was considered from the first day of bleeding until the day of maximum serum LH concentrations and the luteal phase from the next day of serum LH surge until the day before menses began.
LH levels in urine served only as a rough indicator of the onset of ovulation. Blood tests were used to get a more precise indicator of cycle day. The first phase of the cycle, the follicular phase, was considered to be from the first day of menstrual bleeding until the day of the highest level of LH was measured in a blood sample.
In addition, endometrial biopsies were taken from all participants during both control and treated cycles on day LH + 9. This day lies within the implantation window, the time during which the endometrium has optimal receptivity to implantation…
[…]
Endometrial morphology was assessed by correlating the chronological date (day after LH surge) with the morphological endometrial characteristics of specimens as an indicator of hormone action. The dating of the endometrium was related to the serum LH surge, FR, and luteal concentrations of E2 [estrogen] and P4 [progesterone], rather than to the “ideal†28-day cycle, as previously described [22]. The parameters examined were number of glands, stromal edema, and predecidual changes as evaluated by the presence of prominent spiral arteries…
On the ninth day of the luteal phase, a sample of the uterine lining was taken in order to access its receptivity to implantation (had there been an embryo implant).
Timing comparisons between serum with urinary LH demonstrated inconsistencies in 12 out of 90 studied cycles (13.3%). In four control cycles (8.8%) and eight treated cycles (17.7%), urinary LH did not correlate with the day of maximum concentrations of serum LH. In these cases, serum LH, along with E2 and P4 concentrations, were used rather than urinary LH for cycle dating. Thus, eight participants during the treated cycle were identified as not corresponding to the originally assigned groups. Therefore, in four participants originally included in Group B and four in group C, the administration of LNG took place 3 ± 1 day prior to serum LH surge and were reassigned into a new group (group D). This new group received LNG during the late follicular phase, a few days prior to the occurance of LH surge. Thus, the groups studied consisted finally of 15 participants in Group A, 11 in groups B and C, and 8 in Group D.
Like I said, urine LH level is just a rough indicator. As such, it led to premature administration of LNG in 8 cases. Rather than waste data, a fourth study group was formed.
Results
Twelve participants of Group A did not ovulate, and LNG significantly (P<0.05) shortened the mean length of the cycle compared to that of control… These participants were, therefore, excluded from the remainder of the analysis. In the three remaining participants of Group A, LNG administration did not modify significantly the length of that cycle… In Groups B, C, and D, no modifications were noted on cycle length… Follicular phase length was significantly longer only three ovulatory participants of Group A… In the remaining groups, no differences were noted in follicular phase length between treated and control cycles… In all participants of Group B and C, there were no differences in luteal phase length between the treated and control cycles. The three LNG-treated participants with normal cycle length in Group A and all treated participants in Group D had a significant shortening…of the luteal phase.
…12 participants of Group A had anovulatory cycles following LNG…All participants in Groups B, C, and D had ultrasonographic findings of FR, and LNG administration did not modify the day of the cycle at which FR occurred…
Twelve of the fifteen participants in group A (80%) did not ovulate (presumably due to the action of LNG). The remaining three ovulated, but had shorter-than-normal luteal phases. All of the Group B, C, and D participants showed evidence of follicular rupture (and therefore ovulation). Groups B and C showed no change in luteal phase length but Group D had short luteal phases.
With the exception of Group D, the mean serum ILP4–AUC [integrated luteal progesterone area under the curve] in all remaining groups was similar when compared with the control cycle… In Group D, participants presented significantly lower daily serum P4 concentrations and ILP4-AUC when compared with those in control cycles…
That is, while ovulatory cycles in both Groups A and D had shortened luteal phases, Group A had longer follicular phases, but Group D had suppressed progesterone levels during the luteal phase. This is indicative of possible luteal insufficiency in Group D.
In both control and treated cycles, neither inflammatory, reactive, nor other abnormal features in [endometrial] tissue specimens were observed…[E]ndometrial morphology corresponded, according to both LH surge and FR, to the expected day…at which the biopsy was obtained.
[…]
No significant changes were observed between treated and control specimens in any of the studied [endometrial] parameters. No significant differences among any of the groups were observed. Of particular importance was the finding that the predecidual changes as evaluated by the presence of prominent spiral arteries, which are considered crucial for implantation [24], were not altered by LNG.
In short, in this experiment LNG did not seem to make the endometrium hostile to embryo implantation. This is noteworthy because in those cases where LNG shortened the luteal phase, one might expect its normal function would be impaired. If that is the case, it doesn’t seem to be through modification of the endometrium.
Discussion
Results were consistent with other studies showing that preovulatory administration of LNG …suppresses ovulation in most but not all cases…We could not, other wise demonstrate significant alterations in P4 and E2 during the luteal phase when LNG was administered at the time of LH surge or the day after the occurrence of FR. These observations strongly suggest that effects of LNG on the hypothalamic-pituitary-ovarian axis depend on the stage of the menstrual cycle at which the progestin is administered.
We can summarize these results as follows.
- Cycle day 10: 80% didn’t ovulate and 20% ovulated but had shortened luteal phase.
- After LH surge: no observed effect.
- 48 hours after LH surge: no observed effect.
- 2-4 days before LH surge: short luteal phase.
For the cycles in Group A that indicated ovulation, all had shortened luteal phases, but the overall length of each cycle was not significantly different from the control cases. This means that the follicular, i.e., pre-ovulatory, phase was longer than average. Group D was formed because urinary LH was a flawed indicator of LH surge and its use resulted in LNG administration earlier than intended. Coupled with the results of Group A, it seems that if LNG is administered too early, ovulation is not prevented.
In the present study, ovulation occurred in all those women treated immediately before the LH preovulatory surge (Group D); however, in these participants, deficient P4 production with a significantly shorter luteal phase length were observed. Findings in Group A were consistent with an impaired folicular maturation leading to deficient E2 and P4 production during the follicular and luteal phase, respectively…These observations indicate that the preovulatory effects of LNG on the hypothalamic-pituitary-unit are mediated, at least partially, by the progestin’s direct action on the growth, development, and steroidogenic capacity of the ovary to produce adequate E2 concentrations in serum as the primary signal triggering the LH surge…
That is, LNG may act by interfering with follicle development, which leads to deficient estrogen and progesterone production, thereby modifying (shortening, delaying, etc.) the LH surge and shortening the luteal phase.
The finding of ovulatory cycles in three participants belonging to Group A is unexplained but variations in absorption and clearance, as well as differences in ovarian sensitivity after LNG administration, should be considered among the causes of method failure.
They don’t know why some of the Group A participants ovulated, but they offer a suggested route of future inquiry.
On the other hand, the occurrence of ovulation in study participants receiving LNG within 3 days before the onset of LH peak (Group D) may represent either a null-effect or an amplifying P4-like effect of LNG on the hypothalamic-pituitary unit. Under physiological conditions, a small but significant rise in P4 has been considered as the ultimate ovarian signal to trigger gonadotropin preovulatory surge…In addition, P4 administration during follicular phase results in increased amplitude and decreased frequency of LH pulses, which is consistent with the pattern observed during the luteal phase of the cycle… These changes may partially reflect alterations in hypothalamic gonadotropin-releasing hormone secretion that, without apparently affecting preovulatory surge of LH and FR, could be involved in deficient P4 production observed during the luteal phase.
I’m not sure how to interpret this analysis. If I understand correctly, ovulation in Group D participants could mean either that the drug simply had no measurable effect or it acted by modifying the secretion of hormones not studied here, resulting in deficient production of progesterone during the luteal phase. Anyone with greater medical knowledge is encouraged to comment and correct me.
These results suggest that postovulatory contraceptive efficacy of LNG may not involve alterations in the mechanisms associated with endometrial receptivity…
Therefore, it is possible to conclude that interference of LNG with the mechanisms involved in initiating the LH preovulatory surge depends on the stage of follicular development. Thus, anovulation results from disrupting both normal development and hormonal activity of a growing follicle. In addition, the finding that LNG administration at late follicular phase (group D) did not interfere with E2-mediated midcycle gonadotropin surge and ovulation but otherwise did alter P4 production by the corpus luteum requires further investigation, particularly in those at both the ovarian and hypothalamic-pituitary unit, including the interference with preovulatory signals for adequate development and hormonal function of the human corpus luteum.
Since no evidence was found for endometrial effects using LNG, researchers must look to other postovulatory effects if any are be found. One such effect would be luteal insufficiency, i.e., a dysfunctional corpus luteum. The crucial question, in regard to the LNG potentially acting as abortifacient, is whether such luteal insufficiency would result in failed implantation or there would be no embryo to implant (due to some other action of the drug).
Our results may offer a plausible explanation for the contraceptive effects of LNG given postcoitally prior to LH surge or the mechanism involving corpus luteum development. In addition, this study does not support an anti-implantation contraceptive effect of LNG in EC; however, additional targets, besides those described herein, should also be considered and further investigated for the contraceptive effects of LNG.
IOW, these researchers found no evidence to support the hypothesis that LNG acts abortifaciently. However, they studied surgically sterilized women who could not possibly conceive. Therefore, we must be cautious to not apply these results too broadly. Also, though the endometrium was shown to be unaffected by LNG, other postovulatory effects were demonstrated (resulting in shortened luteal phase), leaving open the possibility of anti-implantation effects.
In the next installment, I’ll cover papers that report no postfertilization effects in rats and monkeys given LNG.
References
- Durand M, del Carmen Cravioto M, Raymond EG, Duran-Sanchez O, De la Luz Cruz-Hinojosa M, Castell-Rodriguez A, Schiavon R, Larrea F. On the mechanisms of action of short-term levonorgestrel administration in emergency contraception. Contraception 2001;64(4):227-234.
Pingback: luminousmiseries
Does anyone know whether 45 participants is statistically appropriate to draw generalizations from their findings? I know there’s all kinds of ethical and logistical problems with obtaining more, but I seem to remember from way, way back in pharmacy school tests you could perform to give more credibility to your study by showing the group sizes were appropriate. 45 just seems kind of small to me. Otherwise the study seems well-done.
Fallopian tube dysfunction is another postfertilization effect that has been mentioned. Have you seen anything on that? (And did that recent JAMA review I sent you prove at all useful?)
Stuff: for a fairly involved human study, 45 isn’t necessarily too bad…
I haven’t read anything more current than 2001 about fallopian tube dysfunction. The Croxatto, et al., paper I presented in Part I had this to say:
Pingback: Participatory Bible Study Blog » Blog Archive » Christian Carnival CXLV
45 participants would be too small for a safety or efficacy study, but for a ‘method of operation’ study that is a reasonable number.
Maybe someone with a statistical background can help me out here. I know that any number of test subjects can be statistically significant depending on the desired confidence level one wants in the results. If we had statistics on the number of EC users, would we not be able to calculate the “confidence interval” for 45 test subjects? Thus we would know the confidence with which we can say these results apply to the general population.
Would it not be fair to say that any confidence in these results (> 0%) is better than having no data and hence no confidence? i.e. the conclusions from data presented here is far more likely to be accurate than conclusions based on conjecture and assumptions on the effects of EC?
Based on that study, I would say Plan B does not affect the endometrial lining, but it does affect the length of the luteal phase due to progesterone insufficiency. 80% of those taking the pill before ovulation and before the LH surge failed to ovulate, those who took it after the LH surge or right after ovulation noticed no effect.
Given the pregnancy “prevention” rate of Plan B is 89% (including people who had sex after ovulation), and this study showed a prevention of ovulation of 80%, I would NOT draw the conclusion that Plan B was primarily abortifacient. I would recommend any woman taking Plan B to monitor herself for ovulation and supplement with progesterone to prevent a short luteal phase. I would not recommend women taking Plan B on ovulation day or later, because it is either unnecessary, inneffective, or potentially abortifacient.
To me, it seems that from the study that the primary mode of action of Plan B is to prevent ovulation, not to damage the endometrium (no evidence of this mode of action). It can shorten the luteal phase, but that seems to be due to a crash in hormones, and theoretically could be prevented through progesterone supplementation.
Of course, the results of one small study are not conclusive in any way.
Based on this document: http://www.nccbuscc.org/prolife/issues/abortion/ecfact.htm It would be useful to know exactly when Plan B acts as an abortifacient and when it acts simply as a contraceptive. If the abortifacient action of Plan B is simply to shorten the luteal phase, the combination of a transvaginal ultrasound and progesterone supplements would be a good theoretical protocol to follow to avoid the misuse of the drug in cases of sexual assault.
-Mariana
Please tell me what happened to the women from group A who took Postinor in day 10 of their cycle and still ovulated. I read on your blog that they had normal cycle length, while the luteal fhase was much shorter. Are this data from a study, is it written tehre or is only something you assumed? Thank you.
Day one of my last period was 12/26. I was raped on 1/1 and took plan B on 1/4. Today, 1/11, day 17 of my cycle, I had my period. My cycle is usually about 29 days long, give or take a 2-3 days. I felt none of the warning signs that generally accompany my period. In reading this, I am inclined to think that plan B may be the cause for this unusually short cycle. I will continue to monitor myself for any abnormalities and relate this to my doctor if the need to do so arises.